2025 年 19 巻 1 号 論文ID: oa.2024-0096
Objective: Although vagal paragangliomas (VPs) and carotid body paragangliomas (CBPs) are both neck paragangliomas, they have different surgical risks and clinical courses. In this report, we investigated the feeding arteries of VPs compared with CBPs, with an aim to better differentiate these tumors and improve our understanding of their angioarchitecture.
Methods: We conducted a retrospective analysis of angiography data from 3 cases of VPs and 10 tumors from 9 cases of CBP. For each case, we evaluated the level of the vertebral body corresponding to the upper margin of the tumor, the tumor size, the arterial supply of the tumor, the topological relationship between the external carotid artery and internal carotid artery and the tumor, the details of preoperative embolization, and the incidence of postoperative neurological deficits.
Results: In all 3 cases of VPs, the blood supply originated from the occipital, vertebral, and ascending pharyngeal arteries. By contrast, among the 10 CBP tumors, 3 were supplied by the occipital artery, 1 was supplied by the vertebral artery, and all 10 were fed by the ascending pharyngeal artery. VPs, when compared to CBPs, exhibited larger tumor sizes, a higher positioning of the upper margin of the tumor, and a lack of splaying of the internal and external carotid arteries, compressing both forward. Additionally, preoperative embolization was frequently performed in cases of VPs. Furthermore, the postoperative occurrence of complications such as hoarseness and vocal cord paralysis was also higher.
Conclusion: VPs originate from the inferior ganglion of the vagus nerve, which is chiefly nourished by the vertebral artery. This original arterial distribution may explain the angioarchitecture observed in this study. This study may facilitate the better understanding of the VP angioarchitecture and safe and efficient embolization for them.
Vagal paragangliomas (VPs) are exceedingly rare neoplasms arising from paraganglionic tissue. They comprise 0.6% of all head and neck paragangliomas and just 0.03% of all head and neck tumors. By contrast, carotid body paragangliomas (CBPs) are the most prevalent kind of paragangliomas. Although CBPs are typically benign, VPs have a relatively high malignant potential, with reported rates of up to 19%.1) Consequently, surgical resection is vital for VPs of relatively large size or that exhibit growth, whereas smaller lesions with sluggish growth kinetics may undergo vigilant monitoring.2)
Following surgical intervention for VPs, a relatively high risk of postoperative dysphagia and hoarseness exists,3) which necessitates the disconnection of both the vagus and superior laryngeal nerves. By contrast, CBPs have a greater propensity for vascular complications; these are particularly associated with procedural dissections that involve blood vessels.4) The accurate differentiation between these 2 entities is therefore important for determining the appropriate surgical approach as well as for predicting potential perioperative complications and patient prognoses.3)
Although previous authors have reported the utility of preoperative magnetic resonance imaging, this technique may not consistently discriminate between CBPs and VPs.5) Furthermore, these highly vascularized lesions underscore the importance of preoperative embolization for mitigating intraoperative hemorrhage risk. Accordingly, a sound understanding of the angioarchitecture of these tumors is crucial for executing safe and effective embolization procedures. Although the angioarchitecture of CBPs has been documented,6) that of VPs remains relatively elusive because of their rarity.
The primary objective of the current study was therefore to analyze the detailed angioarchitecture of VPs and compare it with that of CBPs. Our findings will facilitate preoperative differential diagnosis between these entities and help to optimize the embolization process for paragangliomas, thus ensuring enhanced procedural safety and efficacy.
We analyzed 12 cases that underwent angiography at our institution from June 2022 to October 2023, including 3 VPs and 10 CBPs (1 case had bilateral CBPs). All procedures were performed under local anesthesia; procedures included common carotid artery, internal carotid artery (ICA), and external carotid artery (ECA) angiograms as well as selective angiographies via the tributaries of the ECA, depending on cases. To determine tolerance to affected ICA occlusion, a balloon occlusion test was performed in all cases. A definitive diagnosis of VP and CBP was based on intraoperative findings and pathological examinations. Specifically, if a tumor was diagnosed as a paraganglioma in the pathological report and was observed to involve the vagus nerve during surgery, it was diagnosed as a VP. If no involvement of the vagus nerve was observed but there was involvement with the carotid body, a tumor was diagnosed as a CBP. This study was approved by the KEIO University ethics committee (approval no. 20241003), which waived the need for informed consent given the study design under the ethical standards of the 1964 Declaration of Helsinki and its later amendments.
Data collection and analysesWe retrospectively collected the following data from electronic medical records and imaging: patient age, sex, tumor location (laterality and the level of the vertebral body), tumor feeders, whether the common carotid artery was splayed, whether tumor embolization was performed, intraoperative blood loss, surgery duration, resection rate, and postoperative complications. Splaying of the ICA and ECA refers to the separation and displacement of these 2 arteries, which can occur in the presence of certain tumors or pathological conditions in the neck. The angioarchitecture (feeders, drainers, and tumor stain) of each VP and CBP was investigated using conventional 2-dimensional digital subtraction angiography images as well as reconstructed cone-beam computed tomography axial, coronal, sagittal, and oblique images that were reconstructed from 3-dimensional rotational angiography, performed using a C-arm angiography system (Innova 3100; GE Healthcare, Waukesha, WI, USA [until 2018] or Azurion 7 B20/15; Philips, Amsterdam, Netherlands [2019 onward]). The cone-beam computed tomography images were created using a workstation console. Three neurosurgeons performed these investigations (K. Y., K. M, and T. A.). Any discrepancies between the results obtained by the observers were resolved by discussion.
The clinical and radiological findings in the patients with VPs and CBPs are summarized in Tables 1 and 2, respectively. A comparison of radiological findings between the 2 tumor types is shown in Table 3.
| Age (y), sex | Side | Size (mm) | Level | Feeders | CCA splay | Surgical complic. | Embol. | Blood loss (mL) | Resect. rate | Oper. Time (min) |
|---|---|---|---|---|---|---|---|---|---|---|
| 39, M | R | 40 × 27 × 33 | C2 | AphA, OA, VA, ascending cervical artery | No | Vocal cord paralysis | No | 143 | Total removal | 339 |
| 34, F | R | 60 × 41 × 33 | C1 | AphA, OA, VA, PAA, FA or LA | No | Vocal cord paralysis | AphA, PAA | 240 | Partial removal | 365 |
| 35, M | R | 64 × 47 × 36 | C1 | AphA, OA, FA or LA, SThA, VA, deep cervical artery, ECA | No | Vocal cord paralysis, hoarseness | OA, AphA | 377 | Partial removal | 486 |
AphA, ascending pharyngeal artery; CCA, common carotid artery; complic., complications; ECA, external carotid artery; Embol., embolization; F, female; FA or LA, facial or lingual artery; M, male; OA, occipital artery; Oper., operative; PAA, posterior auricular artery; R, right; Resect., resection; SThA, superior thyroid artery; VA, vertebral artery; y, years
| Age (y), sex | Side | Size (mm) | Level | Feeders | CCA splay | Surgical complic. | Embol. | Blood loss (mL) | Resect. rate | Oper. time (min) |
|---|---|---|---|---|---|---|---|---|---|---|
| 59, F | R | 35 × 22 × 17 | C2 | AphA, ECA | Yes | Laryngeal edema | No | 84 | Total removal | 240 |
| 40, F | R | 20 × 20 × 25 | C3 | AphA, ECA | Yes | No | No | 25 | Total removal | 257 |
| L | 50 × 35 × 28 | C2 | AphA, OA, SThA, FA or LA, ECA | Yes | Not yet | Not yet | Not yet | Not yet | Not yet | |
| 37, F | R | 31 × 26 × 19 | C2 | AphA, ECA, SThA | Yes | No | No | 288 | Total removal | 370 |
| 53, F | L | 33 × 25 × 23 | C2 | AphA, ECA | Yes | No | No | 23 | Total removal | 200 |
| 59, F | R | 30 × 21 × 22 | C3 | AphA, ECA, OA, FA or LA | Yes | No | No | 15 | Total removal | 233 |
| 63, F | L | 32 × 23 × 17 | C2 | AphA, FA or LA, ECA | Yes | Dissection | No | 150 | Total removal | 334 |
| 54, F | L | 60 × 32 × 23 | C1 | AphA, OA, FA or LA, SThA, PAA, ECA | Yes | No | Yes | 60 | Total removal | 306 |
| 37, M | R | 44 × 26 × 33 | C2 | AphA, SThA, FA or LA, ECA | Yes | Not yet | Not yet | Not yet | Not yet | Not yet |
| 35, F | L | 73 × 47 × 46 | C1 | VA, OA, MA, FA or LA, ECA | Yes | Not yet | Not yet | Not yet | Not yet | Not yet |
AphA, ascending pharyngeal artery; CCA, common carotid artery; complic., complications; ECA, external carotid artery; Embol., embolization; F, female; FA or LA, facial or lingual artery; L, left; M, male; MA, internal maxillary artery; OA, occipital artery; Oper., operative; PAA, posterior auricular artery; R, right; Resect., resection; SThA, superior thyroid artery; VA, vertebral artery; y, years
| Characteristics | VPs (n = 3) | CBPs (n = 10) | ||
|---|---|---|---|---|
| n | % | n | % | |
| Mean vertical diameter (mm) | 54.7 ± 31.9 | 40.8 ± 11.5 | ||
| Median level of upper margin of tumor | C1 | C2 | ||
| Splaying of carotid bifurcation | 0 | 0 | 10 | 100 |
| Feeder arteries | ||||
| OA | 3 | 100 | 3 | 30 |
| VA | 3 | 100 | 1 | 10 |
| AphA | 3 | 100 | 9 | 90 |
| Direct ECA | 1 | 33.3 | 9 | 90 |
| Surgical complications | ||||
| Vocal cord paralysis | 3 | 100 | 0 | 0 |
| Hoarseness | 1 | 33.3 | 1 | 14 |
| Embolization | 2 | 66.7 | 1 | 14 |
| Mean blood loss (mL) | 253 | 92 | ||
| Mean surgery time (min) | 397 | 277 | ||
| Total removal | 1 | 33 | 7 | 100 |
AphA, ascending pharyngeal artery; CBPs, carotid body paragangliomas; ECA, external carotid artery; OA, occipital artery; VA, vertebral artery; VPs, vagal paragangliomas
Two men and 1 woman had VPs; the patients ranged from 34 to 39 years old. All 3 tumors were on the right side, and in 2 of the tumors, the upper margin of the tumor extended to C1. Tumor sizes ranged from 40 to 64 mm. Representative images are shown in Fig. 1.
One man and 8 women had CBPs; the patients ranged from 35 to 63 years old. Four cases had CBPs on the right side, 4 had CBPs on the left side, and 1 case had tumors on both sides. The upper margin of the tumor extended to C1 in just 2 tumors. Representative images are shown in Fig. 2.
All VPs and CBPs presented with right neck masses but no other neurological deficits. One CBP was a catecholamine-producing tumor; the other CBPs and the VPs were non-catecholamine-producing tumors. Although the VPs appeared to exhibit a larger mean tumor size than the CBPs, this difference was not significant, possibly because of the relatively small number of cases.
AngioarchitectureThe VPs were fed by the occipital artery (OA; n = 3), ascending pharyngeal artery (AphA; n = 3), muscular branch of the vertebral artery (VA; n = 3), deep cervical artery (n = 1), ascending cervical artery (n = 1), facial or lingual artery (n = 2), posterior auricular artery (n = 1), superior thyroid artery (n = 1), and direct small feeders from the ECA (n = 1). The CBPs were fed by the AphA (n = 9), OA (n = 4), VA (n = 1), facial or lingual artery (n = 6), superior thyroid artery (n = 4), internal maxillary artery (n = 1), and direct small feeders from the ECA (n = 10).
In summary, all VPs were supplied by the OA, VA, and AphA, whereas just 4 and 1 of the 10 CBPs were supplied by the OA and VA, respectively, in addition to being supplied by the AphA (n = 9) and ECA (n = 10). Angiographically, all cases of CBPs demonstrated splaying of the ICA and ECA; this was not observed in the 3 cases of VPs, in which both carotid arteries were compressed forward.
Preoperative embolization, surgical procedures, and clinical outcomesIn our institute, we generally conduct preoperative embolization for sizable VPs or CBPs extending to the C1 vertebral level. In these extensive tumors, the upper segments are located beneath the skull base, presenting a challenge because the superiorly approaching feeders cannot be obliterated in the initial surgical phase. In the current study, 2 VPs and 1 CBP were embolized prior to surgery using an Embosphere (Merit Medical Systems, Inc., South Jordan, UT, USA) and coils. For this procedure, a microcatheter was advanced to a point near the tumor, without completely stopping the blood flow. An Embosphere (300–500 or 500–700 μm) was then injected, and was carried along with the blood flow. When the tumor stain weakened, the feeding vessels were occluded using coils. In these cases, the AphAs, posterior auricular arteries, and OAs were embolized. No procedure-related complications were noted after embolization. All VPs and the 7 cases with 7 CBPs were surgically resected. For all VPs, a definite diagnosis was confirmed by vagus nerve involvement; no CBPs had vagus nerve involvement. In each case, the affected vagus nerve from which the VP originated was resected simultaneously with the tumor. Although total removal was achieved in 2 of the 3 VP cases, in 1 case, a partial resection was performed (in which the uppermost part remained) because the tumor extended to the jugular foramen. By contrast, total resection was achieved in all cases of CBPs. Postoperatively, all 3 cases of VPs exhibited right vocal cord paralysis, resulting in hoarseness and swallowing difficulties. Although the vagus nerve was preserved in all cases with CBPs, 1 case had laryngeal edema caused by intubation and another had asymptomatic carotid dissection. Average intraoperative blood loss was 253 mL in the cases with VPs and 92 mL in the cases with CBPs. Average surgery time was 397 min in the cases with VPs and 277 min in the cases with CBPs.
In cases with VPs, postoperative vocal cord paralysis is inevitable because of dissection of the vagus nerve, and perioperative complications are notably higher than in cases with CBPs.3) It is therefore crucial to preoperatively differentiate between CBPs and VPs. On magnetic resonance imaging, 4 distinguishing features of VPs have been reported5): (1) anterior displacement of the ICA without splaying of the carotid bifurcation, (2) involvement of the jugular foramen, (3) relatively large maximum vertical diameters, and (4) an irregular shape. In the 3 cases of VP in this study, no tumors splayed the carotid bifurcation and no cases showed infiltration into the jugular foramen. By contrast, 31% of VPs exhibited splaying of the carotid bifurcation in a previous study.5) It may therefore be difficult to definitively differentiate between VPs and CBPs based on magnetic resonance imaging findings alone.
Angiographically, CBP angioarchitecture has been well described in previous literature.6) However, an angiographical investigation of VPs has yet to be reported. In the current report, we examined VP angioarchitecture and revealed that, compared with CBPs, the involvement of feeders from the OA and VA was frequently observed in VPs; this has not been reported previously. The present findings of VP angioarchitecture, in addition to the aforementioned magnetic resonance imaging features, may therefore facilitate the differential diagnosis between VPs and CBPs.
The present findings will also be useful for performing diagnostic angiographies and planning embolization strategies; the recognition of potential feeders in VPs and CBPs generally leads to a more refined and sophisticated procedure. In the current study, cases with VPs had more extensive blood loss even after embolization; this may be explained by the relatively larger size of VPs and the involvement of multiple feeders from various arteries, including the VA and OA. Hence, to minimize intraoperative bleeding and achieve safer and more effective tumor resection, it is imperative that both interventional radiologists and surgeons are aware of the potential feeders presented in the current study.
Preoperative embolization in VP treatmentParagangliomas are highly vascular tumors; in many cases, embolization is performed before tumor resection to reduce intraoperative bleeding.7) Although there is currently debate about whether embolization should be performed,8) it is considered to have some utility for larger tumors. This is because it can be challenging to secure feeders on the cranial side of the tumor early during resection, and with increasing tumor size, the number of feeding arteries also tends to be higher.6) Some reports have suggested that the vertical diameters of VPs are larger than those of CBPs.5) In our study, the average vertical diameter was 54.7 mm in the VPs and 40.8 mm in the CBPs. The level of the vertebral body corresponding to the upper margin of the tumor was C1–2 in all cases. Additionally, feeder arteries such as the OA, VA, and AphA predominantly supplied blood to the tumors from the cranial side in all cases. In surgical approaches from the neck, the early dissection of feeders from the cranial side is challenging. Preoperative embolization of these vessels is considered effective, and in the current study, embolization was performed in 2 of the 3 cases of VPs. In VPs, as in CBPs, understanding vascular structure through angiography is essential for successful embolization.
Why do we see different angioarchitectures between VPs and CBPs?The reasons underlying the observed different angioarchitectures between VPs and CBPs are intriguing. One possible reason may be related to the arteries supplying the vagus nerve. At the level of the jugular ganglion, where the nerve exits the skull base through the pars venosa of the jugular foramen, arterial supply is provided by the jugular branch of the neuromeningeal trunk of the AphA.9) In the lateral pharyngeal space, below the jugular foramen, the nodose ganglion of the vagus nerve is supplied by collaterals of the ICA and the posterior meningeal artery, which typically originate from the vertebral artery.9) Additionally, it has been reported that the VA anastomoses with the OA extracranially; the most common type is an anastomosis between the VA and the descending branch of the OA.10,11) Moreover, although VPs can arise anywhere along the vagus nerve, they usually arise from the inferior ganglion or plexiform ganglion, which are situated more cranially and more medially than the carotid body.12) Together, these findings indicate that VPs are often supplied by the VA, OA, and AphA (as shown in the schema provided in Fig. 3). By contrast, it has been reported that CBPs are primarily supplied by the AphA because the carotid body itself is supplied by the musculospinal branch of the AphA.6)
Together, our findings suggest that VPs and CBPs may be distinguished from one another by examining their vascular supplies. By accumulating reports on the angioarchitecture of paragangliomas, it may become possible to assess the vascular structure of tumors using non-invasive methods such as computed tomography angiography and magnetic resonance angiography, thus allowing for more accurate tumor differentiation.
We wish to thank Dr. Hiroshi Imagawa for his advice on experimental design.
The authors declare that there are no conflicts of interest related to this work.
